For the company I work for, one of the most important aspects is ensuring we can fallback to different models in case of content filtering since they are not equally sensitive/restrict.

That same argument can be used to say we don't need any technological advancements.

It isn't because some individuals are better fit for something that others wouldn't benefit from help.